When Low Standards are a Winning Strategy: How
Credit Rating Agencies Compete∗
Sean Flynn
Arizona State University
[email protected]
Andra Ghent
Arizona State University
[email protected]
First Draft: February 4, 2014
This Draft: June 3, 2014
Abstract
We empirically analyze the effects of the entry into structured finance products of
new credit rating agencies on the rigor of ratings. Our setting is unique as we study a
period in which the incumbents’ reputation was extremely poor in these assets and the
benefit of more fee income from inflating ratings was low. We nevertheless find that
the entrants issue systematically more generous ratings and that their market share
increases substantially while following this strategy. These results indicate ratings
catering. Furthermore, ratings by incumbents become more generous as the entrants
increase their market share in a product type.
∗
We thank Kim Cornaggia and Dragon Tang for helpful comments on an earlier draft. We gratefully
acknowledge partial funding from the Real Estate Research Institute (RERI) for this project.
1
Introduction
High quality credit ratings can reduce informational asymmetries and transactions costs in
financial markets. Credit ratings provided by a third party can be particularly helpful in
encouraging participation in financial market activities among investors that are less likely
to collect their own information (see Boot and Thakor (1993) for a discussion of market
segmentation by information sensitivity). Conversely, low quality credit ratings can lead to
dysfunction in financial markets. Indeed, Mathis, McAndrews, and Rochet (2009), Ashcraft,
Goldsmith-Pinkham, and Vickery (2010), and Griffin and Tang (2012) have documented
the role of the credit rating agencies (CRAs) in the dysfunction that led to a collapse in
structured finance products in the 2007-2009 period. Ratings were misleading about the
quality of residential mortgage-backed securities (RMBS) in particular and other structured
finance classes more generally. As many investors in structured finance invest primarily in
securities perceived to be information insensitive, and thus rely heavily on credit ratings, the
collapse in trading and issuance of structured products during the crisis is likely related to
investors’ loss of the ability to trust ratings.1
Given the central role that the CRAs play in financial markets, several commentators
including the SEC (2011, 2012) have suggested that one way to improve credit ratings is
to enable greater competition. Indeed, in the spring of 2012, European regulators implemented a framework to increase competition between CRAs (Kanter (2012)). To further
our understanding of how firms compete in the CRA market and the effects of competition
on ratings, we study the entry of two firms into the CRA market. The entrants compete
in ratings for a particular type of structured finance product, commercial mortgage-backed
securities (CMBS). Given the upheaval in the structured finance market in recent years, and
the significant loss of reputation incumbent CRAs suffered in the structured finance market,
there may have been a unique opportunity for a new entrant with rigorous ratings standards
1
See Hanson and Sunderam (2013) for a model of the role of information insensitive securities in the
collapse of securitization markets.
1
to succeed.
We find that the entrants issue systematically higher ratings (often by several notches)
than established CRAs. The entrants’ average ratings are higher than those of each of the
three main incumbents. Moreover, when entrants and incumbents rate the same security,
both entrants rate higher than each of the incumbents, indicating that entrants’ ratings are
not higher due to unobserved heterogeneity in the quality of the securities. The entrants
are much more likely to rate a security AAA than the incumbents, although we find some
evidence that investors perceive a security rated AAA by only an entrant to be riskier than
a security rated AAA by an incumbent.
The entrants’ business strategy of being more lenient appears to be a successful one. By
the end of our sample (March 2014), more than two thirds of all CMBS are rated by at
least one entrant. The entrant that gains significant market share rates more than half of
all CMBS by the end of our sample, which is more than incumbent S&P, while the other
entrant’s market share stagnates at roughly 20% of issues.
Furthermore, using a difference-in-difference approach, we find that the entrants’ more
generous ratings affect the rigor of the incumbents’ ratings. Our main variable of interest is
the entrants’ share of securities ratings in subtypes of CMBS in each year. We simultaneously
control for the year of issuance and the CMBS subtype such that we are not capturing merely
that CMBS ratings became more lax over time or that some subtypes are rated more leniently
than others. We find that as the entrants’ market share increases, the ratings assigned by
incumbent CRAs are more favorable from the perspective of the issuer. A 10 percentage
point increase in the share of securities rated by either entrant raises the average incumbent
rating by about half a grade. As the entrants’ combined market share increases to 68% over
our sample period, this represents an economically meaningful increase in the favorability of
ratings by incumbents. We also find that an increase in the entrants’ share lowers the level
of subordination, a key measure of credit support for structured finance, for securities rated
AAA by at least one incumbent.
2
The effect of new entrants on incumbent ratings is statistically stronger in entrant 2,
whose market share increases substantially over the sample period. Although it does not
enter the market until July 2011, by the end of our sample, entrant 2 rates more than half
of all CMBS and more than one of the three main incumbents. As such, it appears entrant
2 presents a considerably greater competitive threat than entrant 1, whose market share
remains small.
In short, competitive pressures lead entrants to cater to issuers, applying less stringent
criteria to gain business. The presence of rating catering is consistent with the empirical
results of Griffin, Nickerson, and Tang (2013), who, although they do not examine the effect of
entry, find that competition among CRAs leads to ratings inflation in the collateralized debt
obligation (CDO) market. As they emphasize, the Dodd-Frank Act only addresses ratings
inflation that occurs because of explicit rating shopping wherein only the most favorable
ratings are disclosed (as in the theoretical model of Skreta and Veldkamp (2009)), but does
nothing to remedy the conflicts of interest arising from the issuer-pays business model.2 Our
results show that eliminating the ability to shop for ratings is not sufficient for competition
to improve quality.
Our results are related to a number of recent empirical papers on ratings quality. In
particular, they are consistent with those of Becker and Milbourn (2011), who study the
corporate bond market. However, as we explain in the next section, our setting is one in
which competition has a unique opportunity to improve ratings relative to that of Becker
and Milbourn (2011). The finding that competition leads to more lax ratings is almost
certainly a result of the issuer-pays fee scheme used by our sample of CRAs, and our paper
also contributes to the literature documenting the problems with this business model. Jiang,
Stanford, and Xie (2012) find that S&P’s transition from an investor-pay to an issuer-pay
model resulted in higher ratings, and Strobl and Xia (2012) use the investor-paid CRA
Egan-Jones to document that S&P’s ratings are more inflated in situations in which they
2
An issuer-paid CRA generates income from fees it collects from security issuers. In contrast investor-paid
CRAs generate income by charging individual and institutional investors for access to their ratings.
3
face a greater conflict of interest as a result of their issuer-pays business model. Similarly,
Cornaggia and Cornaggia (2013) compare Moody’s with investor-paid Rapid Ratings and
find the latter provides more timely and accurate ratings. Xia (2014) empirically shows that
the entry of an investor-pays CRA improves the quality of ratings.
The remainder of the paper proceeds as follows. The next section explains theoretical
predictions about the effect of competition on the rigor of ratings and relates them to our
unique setting. Section 3 presents our data. Section 4 discusses the ratings of the entrants.
In Section 5, we estimate the effect of entry on the ratings of the incumbents, and Section 6
concludes.
2
Background
2.1
Competition and rating quality: what are the effects and what
are the channels?
That increased competition should lead to worse rating quality is not obvious from either a
theoretical or empirical standpoint. Much of the theory (e.g., Bolton, Freixas, and Shapiro
(2012), Camanho, Deb, and Liu (2012)) suggests that the effect of competition depends on
the reputation of the incumbents. In particular, Camanho, Deb, and Liu (2012) show that
more competition can actually lead to more accurate ratings when the reputations of both the
incumbent and the entrant are low. Intuitively, this occurs because the possibility of gaining
market leadership when reputations are similar is higher than if one CRA has a much better
reputation than the other. Since market leadership is “up for grabs,” both CRAs have an
incentive to rate accurately and make incremental gains in reputation and therefore market
share. Conversely, if reputations are far apart, a “market-sharing” effect dominates, whereby
the CRA with lower reputation will inflate ratings in order to gain additional market share.
Give the unclear theoretical predictions, the effect of competition on ratings is an empirical question, but the empirical results to date have generally been inconclusive. Becker and
4
Milbourn (2011) and Cohen and Manuszak (2013) use data from prior to the financial crisis
and find that increases in Fitch’s market share are associated with more generous credit
ratings. Similarly, Behr, Kisgen, and Taillard (2014) find that rating quality decreased after
the SEC introduced a NRSRO certification process in 1975 that restricted competition. In
contrast, Doherty, Kartasheva, and Phillips (2012) find that when S&P entered the insurance
rating market it actually applied stricter rating standards than the incumbent A.M. Best.3
Even if it is true that competition leads to less stringent ratings, it is still unclear what the
mechanism behind this effect is. Much of the theoretical work (e.g., Skreta and Veldkamp
2009, Bolton, Freixas, and Shapiro 2012, and Sangiorgi and Spatt 2013) has focused on
explicit rating “shopping,” whereby issuers solicit ratings from multiple CRAs and only
purchase the best ones. The presence of shopping does not necessarily indicate that CRAs
are inflating ratings, though: CRAs could be issuing ratings that are perfectly accurate given
their private information, but cross-sectional differences in CRAs’ private information could
lead to differences in disclosed ratings.4
In contrast, rating “catering” arises when CRAs issue ratings that are higher than their
private information dictates for the purpose of garnering more business. Unlike shopping,
catering always implies some degree of rating inflation, and it is therefore a channel that is
distinct from shopping. While Bolton, Freixas, and Shapiro (2012) and Sangiorgi and Spatt
(2013) allow for the possibility of rating catering, to our knowledge only Camanho, Deb, and
Liu (2012) have modeled the effect of competition with catering but with no possibility of
shopping.
3
However, this is likely due to the different incentives insurance companies have to seek additional ratings.
A non-insurance corporate issuer usually seeks additional ratings in order to make its bonds appealing to
investors with “regulatory constraints” (e.g., investors who can only hold bonds with ratings from two or
more CRAs). An insurance company, in contrast, will seek an additional rating only if it allows it to charge
a higher price to buyers of its policies, therefore seeking a more stringent rating is optimal.
4
We use the term rating shopping to refer to the practice of seeking multiple ratings and disclosing only
those ratings that are favorable consistent with the use of the term by Skreta and Veldkamp (2009). It is
worth noting, however, that some empirical work (e.g., An, Deng, Nichols, and Sanders 2014) uses the term
rating shopping to refer to seeking additional ratings even if those ratings are disclosed.
5
2.2
Our Setting
The empirical work on rating competition has yet to identify whether either (or both) of
these mechanisms are at work. The work closest in spirit to our paper, Becker and Milbourn
(2011), is unable to identify whether the deleterious effect they observe that competition has
on rating quality is because of rating catering or rating shopping. Our setup, however, is one
in which competition has the best chance of leading to higher quality ratings. Additionally,
it is one in which we can cleanly show that catering, rather than simply shopping, is a key
channel through which competition leads to more inflated ratings.
First, our data come from a time period and asset class in which the incumbent rating
agencies had very poor reputations. The massive downgrades of billions of dollars of RMBS
and ABS CDOs and the failure of large financial institutions led to public backlash from
lawmakers and lawsuits from investors. As our sample period begins in 2009, we have an
environment in which competition is most likely to lead to more rigorous ratings as predicted
in the model of Camanho, Deb, and Liu (2012).
Second, our setting is one in which the benefit from inflating fee income was small.
Theoretical work (e.g., Bar-Isaac and Shapiro 2013, Bolton, Freixas, and Shapiro 2012)
shows that CRAs are least likely to inflate ratings when the fee income is low. As the
CMBS market has been relatively small post financial crisis, this suggests that the benefits
of inflating ratings to gain business would be low relative to the future benefits of exploiting
a better reputation later. Along this dimension as well, therefore, our setup is one in which
competition has the best chance of improving ratings.
Finally, all of our data come from after the passage of the Dodd-Frank Act, which limits
the ability of security issuers to explicitly shop for ratings. Although the implementation
of Dodd-Frank may not have been complete at the beginning of our sample period, our
results regarding the entrants’ strategies indicate the presence of substantial rating catering.
As such, simply eliminating the ability of issuers to shop for ratings is unlikely to lead to
competition improving the quality of ratings.
6
3
Data
We collect data from Bloomberg terminals on ratings, collateral characteristics, tranche
structure, and coupons of CMBS issued from January 2009 through March 2014. We begin
our sample in 2009 as the disruption in securitization markets resulted in very little issuance
in 2008. Additionally, as we discuss below, securities issued after the financial crisis are quite
different from those issued before.
We include all CMBS except ReREMIC deals, CDOs, or agency multi-family deals.
ReREMICs are more akin to CDOs than traditional CMBS as they are resecuritizations of
existing CMBS tranches. Because they are resecuritizations, they have very different structures from the other CMBS in our sample and Bloomberg does not provide data to control
for the collateral quality in these deals. Furthermore, ReREMICs primarily include securities issued before the financial crisis making them difficult to compare with CMBS backed
exclusively by collateral originated after the financial crisis. Bloomberg usually classifies
multi-family deals backed by the Government Sponsored Entities (GSEs) as collateralized
mortgage obligations (CMOs) such that there are few in our sample to begin with. However,
we drop any deals that have agency-backed flags.
Table 1 summarizes the securities in our sample. Our sample contains 2283 securities
from 253 separate deals. The average security is rated by at least 2 CRAs and some are
rated by 4 CRAs. Moody’s and Fitch each rate more than half the securities. S&P rates a
third and Dominion Bond Ratings Service (DBRS) rates less than a quarter of the securities.
Entrant 1 rates only 363 securities, whereas entrant 2 rates 897, more than S&P. In total,
more than half of the securities are rated by at least one entrant.
We map the alphabetic ratings to a 16 notch numerical scale as follows5 : AAA = 16,
AA+ = 15, AA = 14, AA− = 13, A+ = 12, A = 11, A− = 10, BBB+ = 9, BBB = 8,
BBB− = 7, BB+ = 6, BB = 5, BB− = 4, B+ = 3, B = 2, B− = 1. Almost 50% of
5
The entrants generate ratings on a scale comparable to the incumbents. Hence, the ratings of all six
CRAs (four incumbents plus two entrants) in the sample can be mapped one-to-one to the same numerical
scale.
7
the securities are rated AAA by at least one CRA, and 46.6% are rated AAA by at least
one incumbent, with the remaining 2.7% being rated AAA by only an entrant. The average
rating assigned by incumbents is about one grade lower than the average rating assigned by
the entrants. We discuss in the next section whether the differences in ratings across CRAs
are because of differences in the securities each rates.
Subordination is the main measure of credit enhancement for structured finance products.
It is the percentage of the value of all the securities in the deal that are below it in the priority
of payments. Thus, AAA securities usually have the most subordination and B− tranches
usually have the least. The mean level of subordination of a security in our sample is 19
percentage points. One security in our sample has 75 percentage points of subordination
and some securities have no subordination at all.
The main measure of expected maturity in the CMBS market is the weighted average
life (WAL) which Bloomberg provides in years. The WAL is calculated by projecting the
principal repayment schedule and then calculating the number of years from issuance in
which the average dollar of principal is paid off; see Davidson, Sanders, Wolff, and Ching
(2003) for details. The WAL is calculated under particular assumptions about prepayment
and default, and issuers usually provide a WAL in the prospectus supplement (Bloomberg
populates its WAL field using these supplements). We use this measure to create categories
of maturity: 10% of our securities have an expected maturity of less than 3 years, 10% have
an expected maturity of 3 to 5 years, 10% have an expected maturity of 5 to 7 years, and
the remainder have maturities of 7 years or more.
Previous studies on the effects of rating on yields typically use quarterly or monthly
cross-sectional regressions of the yield or yield spread on rating indicators. A typical framework regresses the bond’s spread to a comparable maturity treasury on a dummy variable
indicating whether the bond is rated by the entrant or on the rating difference between the
entrant and the incumbents. The key feature of these studies (e.g., Kisgen and Strahan
(2010) or Bongaerts, Cremers, and Goetzmann (2012)) is that they use a time series of bond
8
yields and ratings and estimate many cross-sectional regressions.
The inability to access a time series of yields and/or spreads on CMBS makes it impossible
to use such a cross-sectional approach. Reporting requirements for structured products are
much less standardized than for corporate bonds – there is nothing equivalent to TRACE
for these asset classes with the exception of TBA agency securities since May 2011.6 As
such, the vast majority of CMBS do not have current yield or spread information available
in Bloomberg. Bloomberg has transactions prices for some of our tranches on some dates,
primarily the senior tranches. For dates near security issuance, the security prices are extremely close to par so that the spread is a good measure of the return investors expected
to earn.
We thus focus on estimating the effect of CRA entry on the yield at issuance of CMBS.
Specifically, we use the initial coupon spread over comparable maturity Treasuries to proxy
for the yield at issuance. We use the security’s WAL as the security’s maturity and subtract
off the spread on a treasury of comparable maturity in the month the security is issued. The
mean spread is just over 1.9 percentage points and the standard deviation of the spread is
1.0 percentage points.
The securities in our sample vary in the form of the coupons they pay and in their expected
maturity. Floaters, which pay a constant fixed spread to one month LIBOR, comprise 10%
of our sample. An additional 50% are variable rate securities other than floaters, and the
remaining 40% are fixed rate.
Our data contains the shares of each property type backed by the loans in the pool for
the top 3 most common property types in that pool. From the top 3 property type shares,
we construct the shares of retail, office, hospitality, and industrial property. On average,
32% of the loans in a pool are collateralized by retail property, 20% by office property, 15%
6
As of May 2011, the Financial Industry Regulatory Authority (FINRA) requires reporting of all MBS
transactions but has not released the data it has collected for most classes of MBS, including CMBS, to the
public. FINRA has released the data from 2011 onwards to three groups of researchers; see Atanasov and
Merrick (2013), Bessembinder, Maxwell, and Venkataraman (2013), and Hollifield, Neklyudov, and Spatt
(2013). Bloomberg contains modeled prices for many securities but average transactions prices for far fewer
securities.
9
by hospitality property, and 1% by industrial property.
We have three additional variables that describe the collateral, all of which are measured
at origination of the loans. The first variable is the weighted average loan-to-value (waltv )
the mean of which is 60%. Second, we have the weighted average debt-service coverage ratio
(wadscr ) which is the ratio of the net rents (usually called net operating income (NOI)) the
property is expected to earn divided by the required mortgage payment. The mean of wadscr
is 2.2 and the minimum is 1. Third, we have the weighted average maturity (wam) of the
loans backing the security. The mean wam is 97 months consistent with most commercial
mortgages having terms of 7 to 10 years.
The mean issuance date of a security is April 2012. As is well established (see, for example,
Stanton and Wallace (2012)), the CMBS market recovered slowly from the financial crisis.
Thus, issuance of CMBS increases gradually over the sample, with 28 securities issued in
2009, 112 in 2010, 343 in 2011, 550 in 2012, 1006 in 2013, and 244 in the first quarter of
2014.
To account for unobserved heterogeneity in CMBS issuers in some of our empirical analysis, we include the total amount of issuance for the issuer/sponsor (sponsortot) in the year
the security is issued.7 We do so following the finding of He, Qian, and Strahan (2012) for
the RMBS market that larger issuers often get more favorable ratings. The mean security
is issued by an institution with $15 billion of annual CMBS issuance. There is substantial
heterogeneity in annual issuance volume.
CMBS deals differ in their structure and the market is segmented according to the type,
which is important because the CRAs have different methodologies for rating different types.
Our first type is conduit/fusion, which comprise about two thirds of our sample of deals.8
The second category is large loan or single loans, which are deals backed by only a few or one
7
The lead manager is almost always a large financial institution. The issuer is often a SPV ultimately
owned by a large financial institution. We use the prospectuses to identify, to the greatest extent possible,
the ultimate bank sponsor/owner of the SPV.
8
For more information about the institutional structure of the conduit/fusion CMBS market, see An,
Deng, and Gabriel (2011).
10
large loan. We combine the Bloomberg categories Single-Asset and Large Loans into typlarge
since we have relatively few large loan deals that are not only one loan and CRAs usually use
the same methodology for rating Single-Asset and Large Loan deals. Our typlarge category
constitutes 28% of our sample. We group the remaining deals (portfolio, European, and
Small Balance) into an “other” category that contains 5% of the securities in our sample.
Historical Context
To provide a broader context for our estimation sample and to explore differences in securities
issued before and after the crisis, Table 2 summarizes CMBS securities rated by one or more
of Moody’s, S&P, and Fitch during the period 2000-2008.9 As this encompasses the boom
years of 2003-2007, the securities issued during this period display marked differences with
those in our estimation sample. Because DBRS is excluded from the historical sample, we
perform comparisons based only on ratings by Moody’s, S&P, and Fitch. Unreported t-tests
indicate that, with the exception of variable and waltv, all differences between the samples
in Table 2 (historical) and the 2009-2014Q1 sample excluding DBRS ratings are significant
at the 1% level. The number of disclosed ratings is 0.14 grades higher in the 2009-2014Q1
even when we exclude DBRS’ ratings for the latter period.
There are 16,841 securities from 1,298 deals in the historical sample, and the average
security is rated by 2 of the 3 CRAs. The average rating of 10.7 is more than a notch lower
than the average incumbent rating of 11.9 during the estimation period. There is some
variation in average ratings during the historical sample, with the average rising almost
monotonically from 9.7 in 2000 to 11.6 in 2008. On an individual level, S&P issues the
most ratings during 2000-2008, which is in contrast to its last-place position in number of
ratings issued during 2009-2014Q1. It is also the most generous in the historical sample,
although the differences in average ratings between it and the other two incumbents are not
9
Although DBRS was active in the CMBS market during this time, its rating data are not easily accessible
on Bloomberg. However, Moody’s, S&P, and Fitch ratings are representative of ratings during the time
period.
11
economically meaningful.
The average subordination level is nearly 6 percentage points lower in the historical sample. However, it displays significant variation over time, as illustrated in Table 3. Average
subordination levels for all securities and for the set of AAA-rated securities decrease monotonically from 2000 to 2004, but then flatten out from 2005 to 2010 before increasing again
starting in 2011. This pattern could occur for at least two reasons. First, Stanton and
Wallace (2012) report that deal managers began re-tranching the principal balance in AAArated securities in 2004, and that this practice contributed to a sudden spike in subordination
levels for the shortest maturity AAA tranches. Second, the underlying characteristics of the
securities in deals may have changed over time.
To test whether the CRAs exhibit systematic differences over time in how stringently
they rate securities, we estimate
0
Subordinationi,j,t = α0 + γx P eriodDummies + αx0 Controlsi,j,t + i,j,t
(1)
where Subordinationi,j,t is the level of subordination of security j rated by CRA i at time t,
and PeriodDummies are indicators for whether the security was issued during (1) 2000-2003
(firstpd ), (2) 2004-2008 (secondpd ), or (3) 2009-2013 (thirdpd ), respectively. The variables
included as controls are listed in Table 4. Equation 1 is estimated for both the full sample
of securities (columns 1 and 2), and for the AAA-rated subsample only (columns 3 and 4).
The results in Table 4 indicate that the underlying security characteristics explain a large
portion of the changes in subordination levels over time. The variable firstpd is excluded,
so the effects of secondpd and thirdpd are interpreted relative to the period 2000-2003. The
coefficients, which are highly significant with and without robust standard errors, indicate
that moving from the first period to the second increases average subordination by about 0.7
percentage points for all securities and 1.7 percentage points for only AAA-rated securities,
respectively. Moving from the first to the third results in much larger increases of about
12
7.2 and 8.9 percentage points. Robustness checks in which alternative definitions of the
time periods (e.g., 2000-2005, 2006-2008, 2009-2014Q1) are used indicate the effects are
qualitatively similar.
Overall, there are meaningful differences, both statistically and economically, between
the pre- and post-crisis sample of CMBS securities and their ratings. After the financial
crisis, the CRAs required more subordination for securities after controlling for differences
in security characteristics. By beginning our estimation sample in 2009, we avoid variation
caused by changes in the ratings landscape before and after the financial crisis.
4
The Entrants’ Ratings
Both entrants are Nationally Recognized Statistical Rating Organization (NRSROs).10 The
first resulted from the acquisition of a small investor-paid NRSRO by a large investment
advisory services firm that subsequently converted the entity to an issuer-pays model. The
conversion occurred after its acquisition in March 2010 (SEC (2012)), and because we are
interested in studying issuer-paid ratings, we drop the small number of ratings (17 securities
in total) by this entrant prior to its conversion. Entrant 1 also receives revenue from data
services it provides to CMBS investors. Entrant 1 has plans to expand into the RMBS
market and rated its first RMBS deal in late 2013 (Morningstar Credit Ratings, 2013).
Entrant 2’s debut in the CMBS market was January 19th, 2011 (Kroll Bond Ratings,
2011a). This NRSRO, which is more than 40%-owned by pension funds and foundations,
adopted the tagline “[o]ur name is on the line” to underscore its “emphasis on ratings trust
and accuracy” (Kroll Bond Ratings, 2011a). Entrant 2 rated its first deal, a single borrower
transaction, in July 2011 (Kroll Bond Ratings, 2011b). It initially focused only on the large
loan / single asset segment of the market, releasing its methodology for rating such deals on
August 9th, 2011 (Kroll Bond Ratings, 2011c). In 2012, it moved into the conduit/fusion
10
See Beaver, Shakespeare, and Soliman (2006), Kisgen and Strahan (2010), Bongaerts, Cremers, and
Goetzmann (2012), Bruno, Cornaggia, and Cornaggia (2013), and Opp, Opp, and Harris (2013) regarding
the importance of certification for CRAs.
13
market and issued methodology for rating such transactions on February 23, 2012 (Kroll
Bond Ratings, 2012a). By mid-2013 entrant 2 had the third highest market share in CMBS
ratings, and although initially active only in CMBS, it now also rates RMBS, credit card
receivables securitizations, and auto loan securitizations. However, its market share in these
asset classes remains very small.
Reflecting the belief that competition improves the quality of credit ratings, the SEC
permitted both entrants to remain NRSROs, despite them deriving a large share of their
CMBS rating revenue from a handful of issuers, because it was consistent with the SEC’s
goal of enhancing competition (SEC (2011, 2012)). Table 5 documents the evolution of the
entrants’ market share of the CMBS deal types. As Panel A illustrates, Entrant 1 does not
exhibit much forward momentum, rating no securities in 2010 and around 20% in 2011 and
2013.
Entrant 2 enters the market halfway through 2011 such that it rates just 10% of securities that year but 39% of large loan deals, consistent with its initial focus on that market
segment.11 By the first quarter of 2014, it rates 58% of CMBS, giving it the third largest
market share in rating CMBS deals ahead of S&P. Given Entrant 2’s forward momentum in
gaining market share, it likely poses a greater threat to incumbent CRAs than Entrant 1.
As of 2014Q1, Entrant 2 appears to now be focusing on conduit/fusion deals where it rates
two-thirds of all securities issued.
The summary statistics in Table 1 show that both the entrants have higher average ratings
than the three main incumbents. It is possible this occurs because they rate intrinsically
better securities, rather than because their rating methodology is more generous. To explore
this possibility, Table 6 compares the entrants’ ratings to ratings of incumbent CRAs that
rate the same securities. Thus, in Table 6 we hold security characteristics constant, and the
results indicate that both entrants issue systematically more generous ratings of the same
11
The 2% market share of conduit/fusion deals we list in Table 5 is likely because of minor differences
in Bloomberg’s classification of deals relative to the CRAs themselves. We take the Bloomberg deal type
classifications as given to avoid applying our own biases in deal type classifications.
14
security than the main incumbents.
The differences between both entrants’ ratings and those of S&P, Moody’s, and Fitch
are all positive and statistically significant at the 1% level, indicating that the entrants rate
the same security more generously. On average, Entrant 1 rates securities one grade higher
than the three main incumbents, and these differences are statistically significant. There
is no significant difference between Entrant 1’s ratings and those of DBRS. Entrant 2 is
somewhat less generous 1, although on average it still rates a security 0.4 grades higher than
incumbents. The differences between Entrant 2’s ratings and those of Fitch, Moody’s, and
S&P are significant at the 1% level, but there is no significant difference between Entrant
2’s ratings and those of DBRS.
Figures 1 and 2 show that the differences in the ratings between the entrants and the
incumbents are concentrated in the AAA tranches. The figures plot the average rating of the
incumbents against the rating of the entrant for each security rated by both an entrant and
an incumbent. If the entrant’s ratings were similar to the incumbents’, the dots would line
up along the 45 degree line. Consistent with the statistics in Table 6, however, the entrant’s
ratings are often above the 45 degree line in both Figure 1 and Figure 2. In particular,
the entrants frequently issue AAA ratings to securities that the three main incumbents give
AA, A, or even BBB ratings. There are only two instances in which one of the three main
incumbent CRAs issues a AAA rating to a security and the entrant rates it lower than AAA.
The findings in Table 6 and Figures 1 and 2 indicate that the entrants are systematically
more generous, consistent with catering being an important channel through which competition generates rating inflation. If the differences between entrants and incumbents was
simply a result of differences in methodology, we would observe the dots in Figures 1 and 2
randomly scattered around the 45 degree line rather than almost always lying on or above
it.
Given that the entrants are much more likely to issue AAA ratings, a natural question
to ask is whether the market discounts these ratings. To test whether investors treat AAA
15
ratings from entrants and incumbents differently, we estimate
cpnspreadi,j,t = α0 + α1 AAAentrantonlyi,j,t + αx0 Controlsi,j,t + i,j,t
(2)
on the set of securities that are rated AAA by at least one CRA. In equation 2, i indexes the
security, j indicates the deal type, and t indicates the year of issuance. The controls include
dummies for the year of issue, deal type dummies, collateral characteristics, dummies for
the coupon type (fixed rate, floating rate, or variable rate), and the ex ante WAL of the
security in categories. If investors perceive the entrants’ ratings to be a less reliable indicator
of quality than the incumbents’, they will demand a higher return for an issue rated AAA
by only an entrant (AAAentrantonly = 1). A finding that α1 > 0 would thus indicate that
investors do not treat ratings from entrants and incumbents equally.
Column 1 of Table 7 contains the results of estimating (2) on securities of all coupon
types. The coefficient on AAAentrantonly is small and far from statistically significant.
Because the effect of the covariates may differ depending on whether the coupon is fixed
rate, variable, or floating, in column 2 we estimate (2) using only the subset of securities
that have a fixed rate coupon while in column 3 we estimate the model using only securities
that have variable or floating rate coupons. In column 2, the coefficient on AAAentrantonly
is positive and significant at the 10% level, indicating that a security rated AAA by only an
entrant must pay investors roughly 50 basis points more than a security rated AAA by at
least one incumbent. However, given the weak level of significance and the lack of significance
when the sample includes variable and floating rate securities, there is not strong evidence
that investors differentiate between the AAA ratings of incumbents and those of entrants.
As Table 8 illustrates, the differences in the ratings of entrants and incumbents decrease
somewhat as the entrant becomes more entrenched. The table shows the difference between
the entrants’ ratings and those of the incumbents (on the same issue) by year of issuance.
Entrant 1, who makes little progress in gaining market share, provides increasingly more
16
generous ratings relative to those the incumbents provide. In contrast, there is some convergence between Entrant 2’s ratings and those of the incumbents. In 2011, entrant 2’s ratings
were a full grade above the incumbents’ on average. In 2012, the difference between Entrant
2 and the incumbents’ average rating was only 0.5 grades and in 2013 the difference had
shrunk to a mere 0.22 grades. We show in the next section that part of the convergence
is likely due to incumbents becoming less stringent as a result of Entrant 2’s competitive
threat.
5
The Effect of Entry on Incumbent Ratings
As explained in Section 2.1, there are at least two distinct ways in which more competition
can lead to higher disclosed ratings: explicit rating shopping and catering. Since DoddFrank requires issuers to disclose all ratings they solicit, it should have eliminated the former
channel. The potential for catering, though, remains, and entrants issuing systematically
higher ratings would be consistent with this channel. While the entrants appear to give
more generous ratings to gain business, it is unclear whether the incumbent CRAs respond
to the threat. The incumbents may value their reputations sufficiently that they ignore the
competitive pressures, especially because the entrants compete only in structured finance
products, which comprise a small fraction of the incumbents’ overall business.
5.1
Identification Approach
To test whether entry into the CRA market affects incumbents’ ratings, we exploit differences
in the market share of the entrants over time and over subsegments of the CMBS market.
As Table 5 illustrates, there is substantial variation in which types of CMBS the entrants
are active in. For each year and CMBS type, we construct the entrants’ market shares as
the percent of securities they rate. We then include year and CMBS type fixed effects to
control for variation over the business cycle in CMBS ratings and the fact that some CMBS
17
types may be riskier than others.
We estimate
avgratingincumbenti,j,t = α0 + α1 entrant1sharej,t + α2 entrant2sharej,t
+αx0 Controlsi,j,t + i,j,t
(3)
and
avgratingincumbenti,j,t = β0 + β1 entrantsharej,t + βx0 Controlsi,j,t + i,j,t
(4)
where the controls include dummies for the year of issue, deal type dummies, collateral
characteristics, dummies for the coupon type (fixed rate, floating rate, or variable rate), and
the ex ante WAL of the security in categories.
The independent variables of interest are entrant1sharej,t , entrant2sharej,t , and
entrantsharej,t . The variables entrant1sharej,t and entrant2sharej,t are the percentage of
securities of type j issued in year t that are rated by entrants 1 and 2, respectively. The
variable entrantsharej,t is the percentage of securities of type j issued in year t that are
rated by either entrant. Competition results in more generous ratings by the incumbents if
α1 > 0, α2 > 0, or β1 > 0.
The specification implied by (3) and (4) assumes the effect each independent variable
has on incumbent ratings is the same along all notches. This may not be true, however, as
ratings are ordinal in nature. For example, the entrants’ market share may have more of an
effect on whether an incumbent rates a security AA+ vs. AAA than on whether it rates
a security A+ vs. AA-. We thus follow Becker and Milbourn (2011) in estimating (3) and
(4) using both OLS and an ordered probit. The latter preserves the ranking of the different
ratings but does not impose a linearity assumption.
Because a CRA usually rates an entire deal consisting of several securities, rather than
a single security as in the corporate bond market, there is likely to be correlation across
18
the error terms within one deal. As our empirical specifications contain security-specific and
deal-specific variables, we present our results without clustering standard errors and with
standard errors clustered by deal.
Given our finding that most of the entrants’ more lenient ratings are concentrated in the
AAA tranches, as well as the importance of the AAA tranches for issuers more generally,
we also examine whether the entrants altered the tranches that the incumbents rated AAA.
In particular, we are interested in whether the entrants’ propensity to rate tranches AAA
affected the level of subordination of the tranches that the incumbents rated AAA. We
estimate
SubordinationAAA
= α0 + α1 entrant1sharej,t + α2 entrant2sharej,t
i,j,t
+αx0 Controlsi,j,t + i,j,t
(5)
and
0
SubordinationAAA
i,j,t = β0 + β1 entrantsharej,t + βx Controlsi,j,t + i,j,t .
(6)
In estimating (5) and (6), we include only securities that one of the incumbent CRAs rates
AAA.
5.2
5.2.1
Results
Average Incumbent Ratings
Table 9 presents the results from estimating 3 and 4 by OLS. In columns 1 and 3, we include
the individual entrants’ shares. The coefficient on entrant 1’s share is statistically significant
at the 5% level without clustered standard errors, but the coefficient on Entrant 2’s share is
statistically significant at the 1% level when we do not cluster the standard errors by deal
(column 1) and statistically significant at the 5% level when we cluster the standard errors
19
by deal. The economic magnitude of the effect is such that a 10 percentage point increase in
Entrant 2’s market share raises the average incumbents’ rating by 0.27 grades. As Entrant
2 increased its overall market share from 0 to 58% (see Table 5), the effect is economically
important since it implies an increase in average ratings by incumbents of more than one
full grade.
It may appear surprising that Entrant 2’s share has a stronger effect on incumbent ratings
than Entrant 1’s. This is because in Section 3, we found that Entrant 1 had higher ratings
relative to incumbents than Entrant 2, suggesting that it is more prone to inflate ratings.
However, Entrant 1 did not succeed in gaining substantial market share over our period, and
as such, incumbents may not have felt threatened by Entrant 1. Given the sample size, the
statistical insignificance of the coefficient on entrant1share after we cluster standard errors
by deal may be merely due to lack of statistical power rather than any fundamental difference
in the economic influence of the entrants.
The combined effect of the entrants’ shares (columns 2 and 4) has a greater effect on
incumbent ratings than the sum of the individual effects. A 10 percentage point increase
in the share of both entrants combined (i.e., the share of issues rated by either entrant)
raises the average incumbent rating by 0.35 grades. Though Entrant 1’s share alone may
not threaten incumbents, incumbents may feel especially concerned about the competition
when both entrants participate in a market.
As expected, the level of subordination is strongly positively correlated with the average
rating by the incumbents: one additional percentage point is associated with an average
rating 0.33 grades higher. Variable rate securities have lower average ratings than fixed
of floating rate securities. Not surprisingly, securities backed by loans with higher average
LTVs have lower ratings: a 10 percentage point increase in the average LTV results in an
average rating 1.2 grades lower. Securities backed by loans with longer maturities have higher
ratings, and an increase in the weighted average maturity of one year (12 months) increases
the average rating by 0.24 grades. Finally, securities collateralized by a high share of loans
20
backed by hospitality properties have lower ratings. A 10 percentage point increase in the
hospitality share results in an average rating 0.2 grades lower.
Table 10 presents the results from estimating 3 and 4 by ordered probit rather than by
OLS. The results are very similar in character to those from estimating the model via OLS.
5.2.2
Subordination for Securities Rated AAA by an Incumbent
Table 11 contains the results from estimating (5) and (6). Columns 1 and 3 show that, for
tranches that at least one incumbent has rated AAA, a higher market share for Entrant 2
leads to lower subordination. A 10 percentage point increase in the market share of Entrant
2 lowers subordination by 0.8 percentage points. The change in Entrant 2’s overall market
share from the beginning to the end of our sample thus reduced subordination by about four
percentage points. To put this in perspective, the difference in the level of subordination for
a security rated AAA on average vs. one rated AA+ on average is less than three percentage
points. The coefficient on Entrant 2’s share is significant at the 1% level regardless of whether
the standard errors are clustered.12
Consistent with the results in Tables 9 and 10, the results in columns 1 and 3 of Table
11 show that Entrant 1’s share alone does not have a statistically significant effect on the
subordination of AAA tranches. Also consistent with the average rating results, columns 2
and 4 of Table 11 show that the combined entrants’ share has a greater effect on the level
of subordination than the sum of the coefficients on the individual entrants’ shares would
indicate. An increase of 10 percentage points in the share of securities rated by either entrant
results in 1.2 percentage points less subordination.
The WAL of the security has a monotonic effect on the level of subordination: tranches
with longer expected maturities have lower levels of subordination. Variable rate tranches
actually have less subordination than fixed or floating rate tranches. Both of these results
contrast with the effect of WAL and tranche type in the average rating regressions. The LTV
12
The standard errors are less affected by clustering in the models defined by (5) and (6) than in the
models defined by (3) and (3) because there are fewer AAA tranches in deal than total tranches in a deal.
21
of the collateral has a similar effect on the level of subordination as it had in the average
rating regressions. A higher average LTV requires more subordination for a AAA rating just
as a higher average LTV resulted in lower ratings in Tables 9 and 10. The maturity of the
underlying loans variable indicates that securities backed by collateral with longer maturities
require less subordination, consistent with the effect of wam in Tables 9 and 10. However,
the coefficients on wam in Table 11 are never statistically significant.
As in Tables 9 and 10, the results in Table 11 indicate that the incumbent CRAs perceive
the hospitality property sector to be riskier than other property types during our sample
period. In addition, the subordination results show that the CRAs view office property and
retail property as less risky than the omitted property type shares (some non-GSE loans for
multifamily property, health care, mixed use, and property types not disclosed).
6
Conclusions
We have studied the entry of two CRAs on the level of ratings in structured finance. The
entrants issue almost uniformly higher ratings and, in particular, rate many securities AAA
that incumbents do not. The systematically higher ratings of the entrants indicate rating
catering. Furthermore, as the entrants’ market share increases, the incumbents increase their
ratings and lower the level of subordination they provide to AAA tranches. It is too soon
to assess the relative accuracy of the ratings of the incumbents and entrants in our market
given the nature of default in structured finance. However, our results strongly suggest that,
contrary to the stated belief of the SEC and the policy of European regulators, increasing
competition among CRAs is likely to exacerbate, rather than reduce, any tendency the CRAs
have to issue inflated ratings so long as they continue to operate on an issuer-pays model.
22
References
An, Xudong, Yongheng Deng, and Stuart Gabriel, 2011. Asymmetric Information, Adverse
Selection, and the Pricing of CMBS. Journal of Financial Economics, 100, 304-25.
An, Xudong, Yongheng Deng, Joseph B. Nichols, and Anthony B. Sanders, 2014. What
is Subordination About? Credit Risk and Subordination Levels in Commercial Mortgagebacked Securities (CMBS).
Ashcraft, Adam, Goldsmith-Pinkham, Paul and James Vickery, 2010. MBS Ratings and
the Mortgage Credit Boom. FRB of New York Staff Report No. 449.
Atanasov, Vladimir and John J. Merrick Jr., 2013. The Effects of Market Frictions on
Asset Prices: Evidence from Agency MBS. Manuscript, College of William and Mary.
Bar-Isaac, Heski and Joel Shapiro, 2013. Ratings Quality Over the Business Cycle. Journal of Financial Economics, 108, 62-78.
Beaver, William H., Catherine Shakespeare and Mark T. Soliman, 2006. Differential
Properties in the Ratings of Certified versus Non-certified Bond-rating Agencies. Journal of
Accounting and Economics, 42, 303-34.
Becker, Bo and Todd Milbourn, 2011. How did Increased Competition Affect Credit
Ratings? Journal of Financial Economics, 101, 493-514.
Behr, P., Kisgen, D. and J. Taillard, 2014. Did Government Regulations Lower Credit
Rating Quality? Working paper,
SSRN http : //papers.ssrn.com/sol3/papers.cf m?abstracti d = 2412294
23
Bessembinder, Hendrik, William Maxwell, and Kumar Venkataraman, 2013. Introducing
Daylight to Structured Credit Products. Financial Analysts Journal, 69:6, 55-67.
Bolton, Patrick, Xavier Freixas, and Joel Shapiro, 2012. The Credit Ratings Game.
Journal of Finance, 67:1, 85-111.
Bongaerts, Dion, K.J. Martijn Cremers, and William N. Goetzmann, 2012. Tiebreaker:
Certification and Multiple Credit Ratings. Journal of Finance, 67:1, 113-52.
Boot, Arnoud W.A. and Anjan V. Thakor, 1993. Security Design. Journal of Finance,
48:4, 1349-78.
Bruno, Valentina, Jess Cornaggia, and Kimberly J. Cornaggia, 2013. Does Certification
Affect the Information Content of Credit Ratings? Working paper,
SSRN http://dx.doi.org/10.2139/ssrn.1962840
Camanho, Nelson, Pragyan Deb, and Zijun Liu, 2012. Credit Rating and Competition.
Working paper, Catolica Lisbon School of Business & Economics.
Cohen, Andrew and Mark D. Manuszak, 2013. Ratings Competition in the CMBS Market. Journal of Money, Credit and Banking, 45:1, 93-119.
Cornaggia, Jess, and Kimberly J. Cornaggia, 2013. Estimating the Cost of Issuer-Paid
Credit Ratings. Review of Financial Studies, 26:9, 2229-2269.
Davidson, Andrew, Anthony Sanders, Lan-Ling Wolff, and Anne Ching, 2003. Securiti-
24
zation: Structuring and Investment Analysis. Hoboken, NJ: Wiley.
Doherty, Neil A., Anastasia V. Kartasheva, and Richard D. Phillips, 2012. Information
Effect of Entry into Credit Ratings Market: The Case of Insurers’ Ratings. Journal of Financial Economics, 106, 308-30.
Griffin, John M., Jordan Nickerson, and Dragon Yongjun Tang, 2013. Rating Shopping
or Catering? An Examination of the Response to Competitive Pressure for CDO Credit
Ratings. Review of Financial Studies, 26:9, 2270-310.
Griffin, John M. and Dragon Tang, 2012. Did Subjectivity Play a Role in CDO Credit
Ratings? Journal of Finance, 67:4, 1293-1328.
Hanson, Samuel G. and Adi Sunderam, 2013. Are There too Many Safe Securities? Securitization and the Incentives for Information Production. Journal of Financial Economics,
108, 565-84.
He, Jie (Jack), Jun ‘QJ’ Qian, and Philip E. Strahan, 2012. Are All Ratings Created
Equal? The Impact of Issuer Size on the Pricing of Mortgage-Backed Securities. Journal of
Finance, 67:6, 2097-137.
Hollifield, Burton, Artem Neklyudov, and Chester Spatt, 2013. Bid-Ask Spreads and the
Pricing of Securitizations: 144a vs. Registered Securitizations. Manuscript, Carnegie Mellon
University.
Jiang, John (Xuefeng), Mary Harris Stanford, and Yuan Xie, 2012. Does it Matter Who
Pays for Bond Ratings? Historical Evidence. Journal of Financial Economics, 105, 607-21.
25
Kanter, James, 2012. Finance Ministers Clear Way for Credit Rating Competition in
Europe. New York Times, March 31.
Kisgen, Darren J. and Philip E. Strahan, 2010. Do Regulations Based on Credit Ratings
Affect a Firm’s Cost of Capital? Review of Financial Studies, 23, 4324-47.
Kroll Bond Ratings, 2011a. Kroll Bond Ratings Makes Official Debut With Multi-Media
Marketing Campaign. Press Release, Kroll Bond Ratings, January 19, 2011.
Kroll Bond Ratings, 2011b. KBRA Assigns Final Ratings to BAMLL Trust 2011-FSHN.
Press Release, Kroll Bond Ratings, July 14, 2011.
Kroll Bond Ratings, 2011c. Kroll Bond Rating Agency Issues CMBS Single Borrower &
Large Loan Methodology. Press Release, Kroll Bond Ratings, August 9, 2011.
Kroll Bond Ratings, 2012a. U.S. CMBS Multi-Borrower Rating Methodology. Manuscript,
Kroll Bond Ratings, February 23, 2012.
Kroll Bond Ratings, 2012b. Kroll Bond Ratings Agency Rates Inaugural CMBS Conduit
Transaction. Press Release, Kroll Bond Ratings, March 6, 2012.
Mathis, J., J. McAndrews and J.-C. Rochet, 2009. Rating the Raters: Are Reputation
Concerns Powerful Enough to Discipline Rating Agencies? Journal of Monetary Economics,
56, 657-74.
Morningstar Credit Ratings, 2013. Final Ratings Confirmation - Invitation Homes 2013-
26
SFR1. Morningstar Credit Ratings, LLC.
Opp, Christian C., Opp, Marcus M. and Milton Harris, 2013. Rating Agencies in the
Face of Regulation. Journal of Financial Economics, 108, 46-61.
Sangiorgi, Francesco and Chester Spatt, 2013. Opacity, Credit Rating Shopping and
Bias. Working Paper, Carnegie Mellon University.
SEC, 2011. Release No. 34-65339, September 14, 2011.
SEC, 2012. Release No. 34-66514, March 5, 2012.
Skreta, V. and L. Veldkamp, 2009. Ratings Shopping and Asset Complexity: A Theory
of Ratings Inflation. Journal of Monetary Economics, 56, 678-95.
Stanton, R. and N. Wallace, 2012. CMBS Subordination, Rating Inflation, and RegulatoryCapital Arbitrage. Working paper, University of California (Berkeley).
Strobl, Gunter and Han Xia, 2012. The Issuer-Pays Rating Model and Ratings Inflation:
Evidence from Corporate Credit Ratings. Working paper, University of Texas (Dallas).
Xia, Han, 2014. Can Investor-Paid Credit Rating Agencies Improve the Information
Quality of Issuer-Paid Rating Agencies? Journal of Financial Economics, 111, 450-68.
27
Table 1: Summary Statistics
Variable
nratings
numericsp
numericmoodys
numericfitch
numericdbrs
avgratingincumbent
numericentrant1
numericentrant2
avgratingentrant
AAAanyone
AAAincumbent
AAAentrantonly
cpnspread
subordination
floater
variable
walunder3
wal3to5
wal5to7
walover7
retailshare
officeshare
hospshare
indshare
waltv
wadscr
wam
year
sponsortot
typconduitfusion
typlarge
typother
Obs.
2283
770
1508
1348
554
2281
363
897
1177
2283
2283
2283
1868
1700
2283
2283
2006
2006
2006
2006
2143
2143
2143
2143
2152
2088
2214
2283
2283
2283
2283
2283
Mean
2.4
11.7
12.5
12
12.6
11.9
12.7
12.6
12.7
0.499
0.465
0.034
1.955
19.1
0.1
0.5
0.1
0.1
0.1
0.7
32
20
15
1
60
2.2
96.7
2012.4
14,840
0.68
0.27
0.05
28
Std. Dev. Min
0.6
1
4.4
1
4.5
1
4.7
1
4.5
1
4.6
1
4.2
1
4.5
1
4.4
1
0.5
0
0.499
0
0.182
0
1.012
0.005
13
0
0.3
0
0.5
0
0.3
0
0.4
0
0.3
0
0.5
0
28
0
21
0
31
0
4
0
8
8
0.9
1
34.4
12
1.1
2009
10,568
14
0.47
0
0.44
0
0.22
0
Max
4
16
16
16
16
16
16
16
16
1
1
1
8.924
75
1
1
1
1
1
1
100
100
100
28
113
7
540
2014
34,458
1
1
1
Variable definitions in Table 1 are as follows: nratings is the total number of ratings the security received;
numericsp, numericmoodys, numericfitch, numericdbrs, numericentrant1, and numericentrant2 are the
numeric ratings of S&P, Moody’s, Fitch, DBRS, entrant 1, and entrant 2 where 16 corresponds to AAA
and a rating of 1 corresponds to B-. avgratingincumbent is the average rating assigned by the four
incumbent CRAs. avgratingentrant is the average rating assigned by the entrants. AAAanyone takes a
value of 1 if any CRA assigns the security a AAA rating and 0 otherwise. AAAincumbent takes a value of
1 if any incumbent CRA assigns a AAA rating. AAAentrantonly takes a value of 1 if only an entrant CRA
assigns a AAA rating. cpnspread is the annual spread at issuance (in %) that the security pays relative to
a US treasury of comparable maturity. subordination is the level of subordination (in %) of the security.
floater takes a value of 1 if the coupon is a fixed spread above a benchmark index (almost always 1-month
LIBOR). variable takes a value of 1 if the coupon is variable rate other than a floater. walunder3, wal3to5,
wal5to7, walover7 are indicator variables that take a value of 1 if the security’s weighted average life (WAL)
is in the range indicated. retailshare, officeshare, hospshare, indshare capture the percentage of the loans
backed by retail, office, hospitality, and industrial properties. waltv is the weighted average loan-to-value
(LTV) of the loans (in %). wadscr is the weighted average debt service coverage ratio. wam is the weighted
average maturity of the loans measured in months. year is the year of issuance of the security. sponsortot
is the total $ volume (in millions) of CMBS issued by the lead sponsor of the deal in the year the security
is issued. typconduitfusion, largeloan, and typother are indicator variables for CMBS deal types.
29
Table 2: Historical Summary Statistics, 2000-2008
Variable
nratings
numericsp
numericmoodys
numericfitch
avgrating B3
AAA B3
cpnspread
subordination
floater
variable
walunder3
wal3to5
wal5to7
walover7
retailshare
officeshare
hospshare
indshare
waltv
wadscr
wam
year
sponsortot
typconduitfusion
typlarge
typother
Obs.
16841
13150
11301
9582
16841
16841
11939
11440
16841
16841
14065
14065
14065
14065
10922
10922
10922
10922
13637
12016
14682
16841
16841
16841
16841
16841
Mean
2
11.1
11
11
10.7
0.326
1.202
12.4
0.3
0.5
0.2
0.1
0.1
0.6
27
26
5
3
62
1.7
100.4
2004.2
27,289
0.6
0.25
0.16
Std. Dev.
0.5
4.7
4.8
4.7
4.8
0.469
0.763
12.1
0.5
0.5
0.4
0.3
0.3
0.5
23
23
17
8
152
0.6
56.4
2.2
32,809
0.49
0.43
0.36
Min
Max
1
3
1
16
1
16
1
16
1
16
0
1
0
4.109
0
100
0
1
0
1
0
1
0
1
0
1
0
1
0
100
0
100
0
100
0
100
0
7250
1
6
6
529
2000
2008
10
426,300
0
1
0
1
0
1
Notes: 1) avgrating B3 is the average rating assigned by the Big Three incumbents only (Moody’s, S&P,
and Fitch). 2) AAA B3 takes a value of 1 if any one of Moody’s, S&P, or Fitch assign a AAA rating. 3)
See Table 1 for other variable definitions. 4) Although DBRS was actively rating CMBS during the
historical sample, Bloomberg does not have comprehensive information on their ratings such that we focus
on ratings by Moody’s, S&P, and Fitch in our comparision of ratings over our sample period with the
pre-financial crisis period.
30
Table 3: Mean Subordination Levels (%), 2000-2013 (Big Three ratings only)
Year All securities
2000
13.3
2001
13.9
2002
11.3
2003
9.8
2004
9.6
2005
13.1
2006
13.6
2007
13
2008
13.5
2009
13.2
2010
12.9
2011
15.6
2012
18.1
2013
20.9
2014Q1
22.5
31
AAA-rated
28.3
27.4
22.6
18.7
17.5
24.6
27.1
25.8
24.4
21.9
20.5
24.1
28.8
31.5
31
Table 4: Subordination Level Regressions, 2000-2014Q1
secondpd
thirdpd
floater
variable
wal3to5
wal5to7
walover7
waltv
wadscr
wam
retailshare
officeshare
hospshare
indshare
sponsortot
Constant
Year of Issue FEs
Deal Type FEs
SEs Clustered by Deal
Observations
R2
All Securities
(1)
(2)
0.72***
0.72*
(0.27)
(0.43)
7.19***
7.19***
(0.33)
(0.45)
-0.44
-0.44
(0.59)
(1.19)
-5.99***
-5.99***
(0.23)
(0.44)
0.44
0.44
(0.53)
(0.84)
-1.07**
-1.07
(0.54)
(0.88)
-9.79***
-9.79***
(0.48)
(0.77)
0.09***
0.09*
(0.02)
(0.05)
-0.82***
-0.82
(0.27)
(0.58)
0.050***
0.050***
(0.0044)
(0.014)
-0.009*
-0.009
(0.005)
(0.009)
0.0005
0.0005
(0.0049)
(0.0092)
-0.005
-0.005
(0.008)
(0.018)
-0.074***
-0.074***
(0.012)
(0.018)
0.000057*** 0.000057***
(0.000004)
(0.000010)
11.9***
11.9***
(1.4)
(4.0)
No
No
Yes
Yes
No
Yes
9,448
9,448
32%
32%
AAA-rated securities
(3)
(4)
1.68***
1.68***
(0.42)
(0.57)
8.86***
8.86***
(0.47)
(0.55)
0.81
0.81
(0.59)
(0.74)
-0.60**
-0.60
(0.31)
(0.41)
0.63
0.63
(0.48)
(0.45)
-0.62
-0.62
(0.50)
(0.45)
-2.37***
-2.37***
(0.44)
(0.40)
0.17***
0.17***
(0.03)
(0.03)
-2.18***
-2.18**
(0.40)
(0.86)
0.011*
0.011
(0.0065)
(0.013)
-0.030***
-0.030***
(0.007)
(0.012)
0.0065
0.0065
(0.0073)
(0.0123)
0.097***
0.097***
(0.012)
(0.022)
-0.077***
-0.077***
(0.020)
(0.025)
0.000052*** 0.000052***
( 0.000005)
(0.000011)
11.9***
11.9**
(2.3)
(4.8)
No
No
Yes
Yes
No
Yes
3,392
3,392
30%
30%
Notes: 1) Dependent variable is the subordination level of the security. 2) Only securities rated AAA by at
least one incumbent are included in columns 3 and 4. 3) Standard errors are in parentheses. 4)
∗ ∗ ∗p < 0.01, ∗ ∗ p < 0.05, and ∗p < 0.1. 5) Data includes all tranches of CMBS deals issued January 2000
through March 2014 excluding ReREMICS and CDOs. 6) secondpd and thirdpd take values of 1 for the
time periods 2004-2008 and 2009-2014Q1; the omitted issuance year category is 2000-2003. 7) See Table 1
for other variable definitions.
32
Table 5: Share of Securities Rated by Entrants
Year
Panel A: All Deal Types
ratedentrant1
ratedentrant2
ratedentrant
Panel B: Conduit/Fusion Deals
ratedentrant1
ratedentrant2
ratedentrant
Panel C: Large Loan Deals
ratedentrant1
ratedentrant2
ratedentrant
Panel D: Other Deals
ratedentrant1
ratedentrant2
ratedentrant
2009
2010
2011
2012
2013
2014Q1
Total
0%
0%
0%
0%
0%
0%
21%
10%
31%
13%
42%
49%
17%
49%
63%
18%
58%
68%
16%
39%
52%
0%
0%
0%
0%
0%
0%
24%
2%
26%
5%
40%
45%
6%
66%
70%
18%
56%
67%
10%
45%
53%
0%
0%
0%
0%
0%
0%
18%
39%
57%
29%
46%
61%
45%
18%
57%
20%
71%
75%
33%
30%
55%
0%
0%
0%
0%
0%
0%
0%
0%
0%
41%
50%
50%
0%
2%
2%
0%
0%
0%
11%
14%
14%
33
Table 6: Comparison of Entrants’ Ratings with Incumbents’ on Same Issues
Entrant Rating S&P Moody’s Fitch
Panel A: Entrant 1 vs. Incumbents
12.36
11.11
13.29
12.39
13.17
12.52
13.61
12.74
Panel B: Entrant 2 vs. Incumbents
12.54
11.98
13.43
13.13
12.65
12.30
12.99
12.65
Panel C: Entrant Average vs. Incumbents
12.40
11.54
13.42
13.00
12.77
12.36
13.10
12.66
DBRS
Incum. Avg.
difference
N
T-stat
11.73
1.25
0.9
0.65
0.04
1.01
195
166
146
28
363
6.0
5.3
4.8
0.2
8.1
12.24
0.56
0.3
0.35
-0.05
0.41
267
604
527
172
897
4.1
4.7
5.9
-1.5
7.0
12.08
0.86
0.42
0.41
-0.03
0.58
414
743
665
186
1177
6.7
6.6
7.4
-0.9
10.2
13.57
13.04
13.13
34
Table 7: AAA Yields and Securities Rated AAA only by an Entrant
(1)
AAAentrantonly
0.01
(0.22)
subordination
-0.0089***
(0.0031)
floater
-0.06
(0.13)
variable
0.33***
(0.05)
wal3to5
0.38***
(0.05)
wal5to7
0.61***
(0.06)
walover7
0.55***
(0.04)
waltv
0.003
(0.004)
wadscr
-0.11***
(0.03)
wam
0.00052
(0.00126)
retailshare
-0.0039***
(0.0010)
officeshare
-0.0051***
(0.0011)
hospshare
0.0003
(0.0013)
indshare
0.00004
(0.00413)
sponsortot
-1.7e-06
(1.9e-06)
Constant
1.55***
(0.38)
Year of Issue FEs
Yes
Deal Type FEs
Yes
Coupon Type
All
Observations
689
2
R
49%
(2)
0.54*
(0.32)
-0.0088***
(0.0030)
0.55***
(0.04)
0.66***
(0.06)
0.64***
(0.04)
-0.002
(0.005)
-0.16***
(0.05)
0.00003
(0.00116)
-0.0037***
(0.0010)
-0.0047***
(0.0011)
-0.0007
(0.0014)
-0.00032
(0.00371)
-1.1e-06
(1.8e-06)
1.90***
(0.45)
Yes
Yes
Fixed
589
52%
(3)
-0.08
(0.36)
-0.0065
(0.0086)
0.44
(0.50)
-0.63**
(0.26)
0.011
(0.008)
-0.01
(0.06)
0.01231**
(0.00559)
-0.0032
(0.0033)
-0.0040
(0.0031)
0.0017
(0.0029)
0.00169
(0.01928)
-4.4e-06
(5.3e-06)
1.04
(0.88)
Yes
Yes
Floating and Variable
100
72%
Notes: 1) Dependent variable is the spread on the security relative to a US treasury of comparable
maturity. 2) The main variable of interest is AAAentrantonly which takes a value of 1 if only an entrant
rates the security AAA. 3) Standard errors are in parentheses. 4) ∗ ∗ ∗p < 0.01, ∗ ∗ p < 0.05, and ∗p < 0.1.
5) Data includes all tranches of CMBS deals rated AAA by at least one CRA issued January 2009 through
March 2014 excluding ReREMICS and CDOs. 6) See Table 1 for variable definitions.
35
Table 8: Entrant vs. Incumbent Ratings on Same Issues Over Time
Year
Entrant Rating Incumbent Average
Panel A: Entrant 1 vs. Incumbents
2011
12.36
11.93
2012
13.2
12.51
2013
12.77
11.36
2014Q1
12.48
11.52
All Years
12.74
11.73
Panel B: Entrant 2 vs. Incumbents
2011
12.68
11.59
2012
12.6
12.11
2013
12.57
12.35
2014Q1
13
12.27
All Years
12.65
12.24
36
difference
N
T-stat
0.43
0.69
1.41
0.96
1.01
72
74
173
44
363
2.7
2.9
6.8
2.6
8.1
1.09
0.49
0.22
0.73
0.41
34
231
490
142
897
2.3
3.9
4.5
3.3
7
Table 9: OLS Estimation of Effect of Entrants’ Market Shares on Incumbents’ Average
Rating
entrant1share
entrant2share
(1)
3.08**
(1.30)
2.73***
(0.69)
entrantshare
subordination
floater
variable
wal3to5
wal5to7
walover7
waltv
wadscr
wam
retailshare
officeshare
hospshare
indshare
sponsortot
Constant
Year of Issue FEs
Deal Type FEs
SEs Clustered by Deal
Observations
R2
0.33***
(0.01)
0.13
(0.35)
-0.68***
(0.16)
0.49*
(0.26)
0.97***
(0.34)
-0.12
(0.25)
-0.12***
(0.01)
0.27**
(0.11)
0.024***
(0.005)
0.0084*
(0.0045)
0.0057
(0.0048)
-0.021***
(0.005)
-0.021
(0.016)
-3.3e-06
(7.9e-06)
14.2***
(1.4)
Yes
Yes
No
1,567
76%
(2)
3.53***
(0.85)
0.33***
(0.01)
0.14
(0.35)
-0.67***
(0.16)
0.54**
(0.26)
1.00***
(0.33)
-0.09
(0.25)
-0.12***
(0.01)
0.27**
(0.11)
0.025***
(0.005)
0.0087*
(0.0045)
0.0064
(0.0048)
-0.020***
(0.005)
-0.023
(0.016)
-3.9e-06
(7.9e-06)
13.8***
(1.4)
Yes
Yes
No
1,567
76%
(3)
3.08
(2.39)
2.73**
(1.18)
0.33***
(0.01)
0.13
(0.36)
-0.68**
(0.27)
0.49
(0.37)
0.97**
(0.43)
-0.12
(0.27)
-0.12***
(0.03)
0.27
(0.20)
0.024***
(0.008)
0.0084
(0.0059)
0.0057
(0.0068)
-0.021**
(0.008)
-0.021
(0.017)
-3.3e-06
(1.2e-05)
14.2***
(1.9)
Yes
Yes
Yes
1,567
76%
(4)
3.53***
(1.25)
0.33***
(0.01)
0.14
(0.36)
-0.67**
(0.27)
0.54
(0.37)
1.00**
(0.44)
-0.09
(0.26)
-0.12***
(0.03)
0.27
(0.19)
0.025***
(0.008)
0.0087
(0.0059)
0.0064
(0.0069)
-0.020**
(0.008)
-0.023
(0.017)
-3.9e-06
(1.2e-05)
13.8***
(1.9)
Yes
Yes
Yes
1,567
76%
Notes: 1) Dependent variable is the average rating of the security by incumbent CRAs. 2) Standard errors
are in parentheses. 3) ∗ ∗ ∗p < 0.01, ∗ ∗ p < 0.05, and ∗p < 0.1. 4) Data includes all tranches of CMBS
deals issued January 2009 through March 2014 excluding ReREMICS and CDOs. 5) See Table 1 for
variable definitions.
37
Table 10: Ordered Probit Estimation of Effect of Entrants’ Market Shares on Incumbents’
Average Rating
entrant1share
entrant2share
(1)
1.44**
(0.66)
1.31***
(0.36)
entrantshare
subordination
floater
variable
wal3to5
wal5to7
walover7
waltv
wadscr
wam
retailshare
officeshare
hospshare
indshare
sponsortot
Year of Issue FEs
Deal Type FEs
SEs Clustered by Deal
Observations
Pseudo-R2
0.17***
(0.01)
0.12
(0.22)
-0.36***
(0.08)
0.39**
(0.19)
0.37
(0.26)
-0.76***
(0.23)
-0.064***
(0.0070)
0.16***
(0.06)
0.019***
(0.003)
0.0045*
(0.0024)
0.0038
(0.0025)
-0.0099***
(0.0025)
-0.012
(0.008)
-1.6e-06
(4.1e-06)
Yes
Yes
No
1,567
32%
(2)
1.65***
(0.43)
0.17***
(0.01)
0.13
(0.22)
-0.36***
(0.08)
0.43**
(0.18)
0.39
(0.25)
-0.74***
(0.22)
-0.065***
(0.0070)
0.16***
(0.06)
0.019***
(0.003)
0.0047**
(0.0024)
0.0041
(0.0025)
-0.0099***
(0.0024)
-0.012
(0.008)
-1.8e-06
(4.1e-06)
Yes
Yes
No
1,567
32%
(3)
1.44
(1.33)
1.31**
(0.65)
0.17***
(0.01)
0.12
(0.33)
-0.36***
(0.12)
0.39
(0.39)
0.37
(0.50)
-0.76**
(0.35)
-0.064***
(0.017)
0.16*
(0.09)
0.019***
(0.004)
0.0045
(0.0032)
0.0038
(0.0036)
-0.0099**
(0.0047)
-0.012
(0.009)
-1.6e-06
(6.9e-06)
Yes
Yes
Yes
1,567
32%
(4)
1.65**
(0.69)
0.17***
(0.01)
0.13
(0.33)
-0.36***
(0.12)
0.43
(0.38)
0.39
(0.49)
-0.74**
(0.34)
-0.065***
(0.016)
0.16*
(0.09)
0.019***
(0.004)
0.0047
(0.0032)
0.0041
(0.0036)
-0.0099**
(0.0046)
-0.012
(0.009)
-1.8e-06
(6.9e-06)
Yes
Yes
Yes
1,567
32%
Notes: 1) Dependent variable is the average rating of the security by incumbent CRAs. 2) Standard errors
are in parentheses. 3) ∗ ∗ ∗p < 0.01, ∗ ∗ p < 0.05, and ∗p < 0.1. 4) Data includes all tranches of CMBS
deals issued January 2009 through March 2014 excluding ReREMICS and CDOs. 5) See Table 1 for
variable definitions.
38
Table 11: Entrants’ Market Shares and Subordination of Tranches Rated AAA by an Incumbent
entrant1share
entrant2share
(1)
-6.50
(4.15)
-7.89***
(2.24)
entrantshare
floater
variable
wal3to5
wal5to7
walover7
waltv
wadscr
wam
retailshare
officeshare
hospshare
indshare
sponsortot
Constant
Year of Issue FEs
Deal Type FEs
SEs Clustered by Deal
Observations
R2
1.38*
(0.83)
-2.34***
(0.61)
-0.36
(0.59)
-1.79**
(0.75)
-1.98***
(0.52)
0.34***
(0.05)
-0.51
(0.40)
-0.018
(0.014)
-0.030**
(0.013)
-0.038***
(0.014)
0.099***
(0.015)
-0.021
(0.049)
0.000071***
(0.000023)
-0.18
(4.83)
Yes
Yes
No
713
65%
(2)
-12.0***
(2.8)
1.30
(0.82)
-2.36***
(0.61)
-0.40
(0.58)
-1.82**
(0.75)
-1.99***
(0.52)
0.34***
(0.05)
-0.56
(0.40)
-0.020
(0.014)
-0.032**
(0.013)
-0.041***
(0.014)
0.097***
(0.015)
-0.022
(0.049)
0.000074***
(0.000023)
0.85
(4.76)
Yes
Yes
No
713
65%
(3)
-6.50
(6.01)
-7.89***
(2.82)
1.38
(0.91)
-2.34***
(0.63)
-0.36
(0.47)
-1.79**
(0.69)
-1.98***
(0.30)
0.34***
(0.09)
-0.51
(0.63)
-0.018
(0.024)
-0.030
(0.018)
-0.038
(0.023)
0.099***
(0.024)
-0.021
(0.062)
0.000071
(0.000043)
-0.18
(7.34)
Yes
Yes
Yes
713
65%
(4)
-12.0***
(3.9)
1.30
(0.91)
-2.36***
(0.62)
-0.40
(0.47)
-1.82***
(0.70)
-1.99***
(0.30)
0.34***
(0.09)
-0.56
(0.62)
-0.020
(0.023)
-0.032*
(0.018)
-0.041*
(0.023)
0.097***
(0.023)
-0.022
(0.061)
0.000074*
(0.000044)
0.85
(6.92)
Yes
Yes
Yes
713
65%
Notes: 1) Dependent variable is the subordination level of the security. 2) Only securities rated AAA by at
least one incumbent are included. 3) Standard errors are in parentheses. 4) ∗ ∗ ∗p < 0.01, ∗ ∗ p < 0.05, and
∗p < 0.1. 5) Data includes all tranches of CMBS deals issued January 2009 through March 2014 excluding
ReREMICS and CDOs. 6) See Table 1 for variable definitions.
39
Figure 1: Entrant 1 vs. Other CRA’s Ratings
10
5
0
numericentrant1
15
Numeric Ratings: 16=AAA, 1=B-
0
5
10
15
numericsp
10
5
0
numericentrant1
15
(a) S&P
0
5
10
15
numericmoodys
10
5
0
numericentrant1
15
(b) Moody’s
0
5
10
numericfitch
(c) Fitch
40
15
Figure 2: Entrant 2 vs. Other CRA’s Ratings
10
5
0
numericentrant2
15
Numeric Ratings: 16=AAA, 1=B-
0
5
10
15
numericsp
10
5
0
numericentrant2
15
(a) S&P
0
5
10
15
numericmoodys
10
5
0
numericentrant2
15
(b) Moody’s
0
5
10
numericfitch
(c) Fitch
41
15